As organisations continue on the path to becoming more data-driven, data governance must become a priority on corporate agendas. In simple terms, data governance refers to the aspects an organisation must put in place to ensure data quality, security and compliance. Julian Thomas, Principal Consultant at PBT Group, notes that to implement data governance effectively, an organisation must first define what their requirements for each of these items are and must view data governance as non-negotiable.
“In my view, security is probably the easiest to define requirements for, as the questions that need to be asked are generally easy to understand and answer. Data quality is trickier as one must ponder what the acceptable levels of data quality are for the different areas of the business. Lastly, data compliance is probably the most difficult. Compliance normally comes in two parts: internal rules that the company has implemented, and rules enforced by external regulatory bodies or by government,” says Thomas.
In all these scenarios, it is the role of data architects and data governance specialists to define and document the various rules around these aspects, and what is expected of implementation teams in terms of adherence to these rules.
“Despite this, in my experience, a great deal of responsibility is often placed on the shoulders of a small number of people in the business to document these rules. This takes a great deal of time, and the result is generally large, complex documents that are shared with the organisation periodically,” he says.
The problem is this is only socialised at certain management levels of the organisation. “Very often, the people who implement things in the business are not aware of the details behind this. In many cases, implementing data governance is left to the end of a project, as a final check box. This results in a scramble when someone raises the governance aspect as a requirement, and the implementation team is left, in the 11th hour, to come up with a response to data governance requirements,” Thomas stresses.
Frequently, data governance is not addressed as part of an implementation, and it is only routine audits that identify this resulting lack of compliance.
“I believe that solution architects, system designers and business analysts must treat data governance as a critical, non-negotiable feature deliverable. Only by proactively listing data governance as a feature and prioritising this upfront from the outset of a project, will we see real implementation of data governance that adds value,” he says.
Data governance must therefore be “given a seat at the table” from the earliest phase of the project and be included as a key deliverable in business requirements definition. Therefore, it must be part of the solution architecture and design. This will result in prioritised activities in the project planning.
“To accomplish this, the overall mindset of an organisation must change. Instead of seeing data governance as a burden that requires the least effort to implement, it must be viewed as something that adds value to the business and is treated as a critical requirement.
“Also, the way data governance is implemented must change. It must be understood to be critical and therefore needs to be implemented with the same level of zeal as we would any other critical business feature,” he concludes.
Share